3.1. Availability of Zinc and Manganese in Soils
Soil constitutes a specific environment that constantly undergoes spatial and temporal variations. The processes occurring in this habitat are interrelated, interact with each other and could be a result of the influence of natural and anthropogenic factors.
On the day of the waste sulfur and organic material application, the zinc content in very light soil ranged from 0.388 to 0.808 mg Zn kg
−1 d.m., and in heavy soil, it varied from 2.02 to 3.03 mg Zn kg
−1 d.m. (
Table 4). During incubation, the zinc content in both soils increased. One hundred and twenty days after the materials were introduced, the zinc content in the very light and heavy soils amounted to 1.628 to 2.933 mg Zn kg
−1 d.m. and 3.73 to 4.95 mg Zn kg
−1 d.m., respectively (
Table 3).
After 120 days of incubation, the zinc content in the treatments of very light and heavy soils with the addition of waste sulfur and its mixtures with organic materials was, as a rule, comparable to or significantly higher than that in the control treatment (without additives). Among the treatments with the applied materials in very light soil, the treatment with the addition of waste sulfur at the SI sulfur dose and biochar and the treatment with the addition of waste sulfur at the SII sulfur dose and digestate were characterized by significantly higher zinc content, while treatments with the lowest zinc were waste sulfur at the SI and SII sulfur doses and the mixture of waste sulfur at the SII sulfur dose with manure. Regarding the zinc concentration in heavy soil treatments at the end of the experimental period, treatments with the addition of waste sulfur at the SII sulfur dose and the mixture of waste sulfur at the SI sulfur dose with digestate had the highest zinc content. Statistically significant differences among other treatments with the addition of the tested materials were relatively small and related only to some objects.
At the end of the experiment, a slight effect of the sulfur dose on the zinc concentration of both soils tested (regardless of organic material addition) was observed. Throughout the incubation period, the content of the discussed element in treatments with the addition of waste sulfur at the SII sulfur dose and its mixtures with organic materials was comparable to or significantly higher than the content of this element determined in the treatments with the addition of waste sulfur at the SI sulfur dose and its mixtures with organic materials.
In relation to treatments with the addition of only waste sulfur, a beneficial effect of the introduced organic materials (manure, biochar and digestate) on zinc concentration in both tested soils was observed (especially in the treatments of very light soil).
On the day of the waste sulfur and organic material application, the manganese content in very light soil ranged from 2.99 to 4.86 mg Mn kg
−1 d.m., and in heavy soil, it varied from 5.00 to 6.92 mg Mn kg
−1 d.m. (
Table 4). During incubation, the manganese content in both soils increased. One hundred and twenty days after the materials were introduced, the manganese content in the very light and heavy soils amounted to 4.93 to 6.38 mg Mn kg
−1 d.m. and 13.18 to 14.26 mg Mn kg
−1 d.m., respectively (
Table 5).
After 120 days of incubation, the manganese content in the treatments of very light and heavy soils with the addition of waste sulfur and its mixtures with organic materials was comparable to or significantly higher than that in the control treatment (without additions). Among the treatments with the applied materials in very light soil, the treatment with the addition of waste sulfur at the SI sulfur dose and biochar and the treatment with the addition of waste sulfur at the SII sulfur dose and digestate were characterized by the highest manganese content. Statistically significant differences among other treatments with the addition of the tested materials were relatively minor and related only to some objects. Regarding the manganese concentration in heavy soil treatments after 120 days of the experiment, statistically significant differences among treatments with the addition of the tested materials were relatively small and related only to some objects. However, all determined values of the studied element content were significantly higher than those determined in the control treatment.
At the end of the experiment, a minor effect of the sulfur dose on the manganese concentration of both tested soils (regardless of organic material addition) was observed. Throughout the incubation period, the content of the discussed element in treatments with the addition of waste sulfur at the SII sulfur dose and its mixtures with organic materials was, as a rule, comparable to or significantly higher than the content of this element determined in the treatments with the addition of waste sulfur at the SI sulfur dose and its mixtures with organic materials.
In relation to the treatments with the addition of only waste sulfur, a beneficial effect of the introduced organic materials (manure, biochar and digestate) on the manganese concentration in both tested soils was observed.
The presented findings indicate that the application of waste sulfur and its mixtures with organic materials affected the zinc and manganese contents of both tested soils. After 120 days from the introduction of the tested materials, in comparison to the control treatment, the contents of discussed elements in the very light and heavy soils increased. Similar results have been previously reported; e.g., Soaud et al. [
46], after the application of elemental sulfur at a dose of 444 mg kg
−1 of soil, found that the concentrations of Mn and Zn increased (in comparison to the control treatment) in all three tested soils during a 128-day incubation experiment. At the end of the incubation period, the zinc concentration after amending the soil with elemental sulfur ranged from 0.21 to 1.93 mg kg
−1, while in the control treatment, the determined element content amounted to 0.34 to 1.77 mg kg
−1. As a rule, the zinc concentration tended to decrease during the incubation period. After 128 days of the experiment, treatments with elemental sulfur were characterized by manganese contents ranging from 2.89 to 4.47 mg kg
−1, while control treatments were characterized by this element content ranging from 0.23 to 1.00 mg kg
−1. Cifuentes and Lindemann [
47] found that after the introduction of elemental sulfur (at a dose of 5 g S
0 kg
−1 soil), the available Mn content had increased significantly by the end of the 270-day field experiment, while this treatment had no effect on other nutrients (P, Fe and Zn). The addition of organic matter (fresh manure, composted manure and the remains of bermudagrass) increased the contents of P and Zn. The authors highlighted that the increase in elemental sulfur oxidation resulting from the organic matter added to S
0-treated soils significantly increased Mn availability only (and had no effect on P, Fe or Zn) in comparison to treatments with the addition of only S
0 or organic matter. The lack of an increase in the availability of P, Fe or Zn after the application of only S
0 could be the result of an insufficient sulfur dose and its slow and incomplete oxidation or a high soil buffering capacity, which resists soil acidification. Furthermore, elements could have been immobilized by soil microorganisms, chemically reprecipitated in less mobile forms or not affected at all [
47]. Skwierawska et al. [
48] found that after the application of S
0 at annual doses of 40 kg ha
−1, 80 kg ha
−1 and 120 kg ha
−1, the Zn and Mn concentrations in the soil after a three-year field experiment amounted to 3.69, 3.30 and 3.99 mg Zn kg
−1 soil, respectively (2.92 mg Zn kg
−1 soil in the control treatment), and 93.19, 90,16 and 91.80 mg Mn kg
−1 soil, respectively (85.75 mg Mn kg
−1 soil in the control treatment). The depletion of these element resources could have resulted from their enhanced solubility in the soil solution and their utilization by plants. The opposite results were presented by Cui and Wang [
49], who found a decrease in the available Zn concentration after soil fertilization with elemental sulfur, while Tabak et al. [
50] stated that significant variations in the concentrations of available forms of heavy metals (Fe, Mn, Zn, Cu, Cr, Ni, Pb and Cd) as a result of elemental sulfur application were found only for some of the determined elements and used reagents (0.01 mol L
−1 CaCl
2, Mehlich 3, 1 mol L
−1 HCl). This can be explained by the slight impact of sulfur introduction on soil pH and the low contents of elements (constituting impurities) other than sulfur in the applied waste.
Introduced elemental sulfur undergoes biological transformation into ions (SO
42−), constituting an available form of this element [
51]. Together with sulfur oxidation in soil, sulfate ions are produced, and hence, soil acidification is induced [
52,
53]. There are many factors that affect the solubility and bioavailability of nutrients and trace elements; however, the most crucial is the value of soil pH. Any factor affecting the pH value affects other soil features by shaping its reaction. As a rule, the contents of the available forms of macroelements decrease with decreasing pH values. This is especially visible in the case of phosphorus, the availability of which can be reduced by up to 90% in an acidic environment [
54]. On the other hand, as the pH value decreases, the solubility of trace elements increases. At lower pH values, the mobility of cations is enhanced, while the activity of anions is reduced. As the pH value increases, the electromagnetic charge of soil colloids (organic matter, iron and aluminum oxides, and 1:1 and 2:1 clay minerals) also increases, as a result of which cation immobilization and anion mobility are higher. Furthermore, together with the increasing soil reaction, metal precipitation into carbonates, chlorides, hydroxides, phosphates and sulfates also increases.
Both manganese and zinc are essential nutrients involved in physiological processes occurring during plant growth and development [
47,
55,
56,
57]. Manganese participates in water photolysis in chloroplasts, the regulation of enzyme activities and protection from oxidative membrane damage, while zinc constitutes an element of some enzymes, maintains membrane integrity, and regulates auxin synthesis and pollen production [
58] based on various sources. Although the availability of these elements constitutes an important factor in increasing crop yield quality and quantity, their toxic nature could be revealed under acidic soil conditions [
55,
59,
60].
3.2. Regression Models to Predict Changes in Manganese and Zinc Availability over Time
The results of the regression analysis of the Mn availability changes over time in very light soil for particular experimental treatments are presented in
Figure 1. However, out of nine treatments, only two regression models turned out to be statistically significant: SI + B (soil with the addition of waste sulfur (sulfur dose: I) and biochar) and SII + D (soil with the addition of waste sulfur (sulfur dose: II) and digestate). The coefficients of determination (R
2) for the developed regression equations ranged from 0.9139 (SI + B) to 0.8073 (SII + D), and the adjusted coefficients of determination (R
2adj.) reached values from 0.8852 to 0.7430, respectively. The coefficient of determination for other treatments ranged from 0.0974 to 0.7434.
For all regression models developed for Zn in very light soil in all experimental treatments, the
p values were below 0.05 (
Figure 2). The results showed that in the control treatment, the prediction of Zn availability in very light soil was the most precise (R
2 = 0.9842, R
2adj. = 0.9790). In other treatments, the regression equation developed for the combination of waste sulfur (sulfur dose: I) and biochar (SI + B) could also predict Zn with very high accuracy (R
2 = 0.9706, R
2adj. = 0.9608). In other cases, R
2 ranged from 0.7962 (SII + M) to 0.9611 (SII + D).
As shown in
Figure 3, only two simple regression models of Mn changes in heavy soil were statistically significant, and both of them were amended with waste sulfur at dose I and dose II (SI and SII, respectively). The coefficients of determination (R
2) for the presented regression equations ranged from 0.8012 (SI) to 0.7804 (SII), and the adjusted coefficients of determination (R
2adj.) reached values from 0.7350 to 0.7071, respectively. Other linear models were less accurate: R
2 ranged from 0.5934 to 0.7287, and
p values were below 0.05.
Contrary to the observed statistical significance of the models of Zn in all treatments for very light soil, for heavy soil, only four regression models had
p values lower than 0.05 (
Figure 4). In heavy soil, Zn changes over time were predicted the most accurately in the following treatments: SII (soil with the addition of waste sulfur (sulfur dose: II); R
2 = 0.9020, R
2adj. = 0.8693), SI + M (waste sulfur (sulfur dose: I) and manure; R
2 = 0.8229, R
2adj. = 0.7639), SI + D (waste sulfur (sulfur dose: I) and digestate; R
2 = 0.8250, R
2adj. = 0.7667) and SI + B (waste sulfur (sulfur dose: I) and biochar; R
2 = 0.8235, R
2adj. = 0.7647). In other cases, the models were not significant (
p > 0.05), and R
2 ranged from 0.4681 to 0.8777. An additional statistical analysis allowed the authors to develop significant polynomial regression models for SI (soil with the addition of waste sulfur (sulfur dose: I); R
2 = 0.9850, R
2adj. = 0.9699;
) and for SII + M (waste sulfur (sulfur dose: II) and manure; R
2 = 0.9683, R
2adj. = 0.9365;
). Linear models were not statistically significant for these experimental treatments.
Figure 5 shows the observed Mn and Zn contents plotted against the fitted values calculated on the basis of the obtained regression equations, indicating how well the models describe the changes in the availability of these elements in the soil in a particular experimental treatment (SI + B: soil with the addition of waste sulfur (sulfur dose: I) and biochar). Despite the high coefficients of determination for both presented models, a clearly better fit of the data can be observed for the Zn model.
A similar procedure for plotting observed versus predicted data was used for the chosen time models developed for heavy soil (
Figure 6). Treatment SI for Mn and treatment SII for Zn were selected according to their statistical significance. In addition, the highest possible value of the coefficient of determination R
2 was chosen, which was 0.8012 for Mn (SI) and 0.9020 for Zn (SII). For heavy soil data, the data points for the Zn model are much closer to the projected regression line, and it can be concluded that the regression model fits the data reasonably well and much better than that of Mn.
3.3. Application of Multiple Regression Analysis for Predicting Manganese and Zinc Availability in Soils
Multiple linear regression analysis was used to develop models that describe how the y variable (Mn and Zn) relates to a number of explanatory variables (x
n: soil pH value, S-SO
4 and dehydrogenase activity). The general structure of the base regression equation for particular elements used for the analysis was:
, where b
0 is the intercept; b
1 … b
n are regression coefficients (not standardized); and x
1, x
2 and x
3 are the soil pH value, S-SO
4 content and dehydrogenase (D) activity, respectively. The models were developed on the basis of data from five incubation periods (days 0, 15, 30, 60 and 120). Statistical analyses were performed separately for very light and heavy soils. After stepwise regression analysis with backward elimination, the final and simplified models are presented in
Table 6.
In the regression equations after the backward elimination procedure, the explanatory variables were often not the same. Of the four regression models, there was a variable related to soil pH in the three equations: for Zn content in both very light and heavy soils and for Mn content in heavy soil; in the case of the regression equation for Zn content in heavy soil, the sulfate sulfur content turned out to be an additional variable remaining after the analysis (
Table 6). Dehydrogenase activity only remained in one regression equation, which was developed for Mn content in very light soil. The coefficient of determination was the highest (R
2 = 0.8409), and S
e was the lowest (0.3164) for the model describing Zn content in heavy soil.
It can be seen that in
Figure 7a,b, the observed data and their adjustment to the regression line are rather dispersed. As mentioned before, the only case when the variable indicating the activity of dehydrogenase remained in the regression equation was the model describing the availability of Mn in very light soil. Despite the fact that the statistical analysis proved the significance of the model, the coefficient of determination was very low (R
2 = 0.1934), indicating the poor fit and precision of the model. For a given model, there was higher variability around the regression line (
Figure 7a), which is the reason for its lower R
2 value. In very light soil, the regression equation for Zn consisted of pH, but the accuracy of the model was much lower compared to that obtained for Zn in heavy soil, as evidenced in
Table 5 by R
2 (0.4257) and S
e (0.5072), together with the visual presentation in
Figure 7b. The equation for Mn in heavy soil (
Figure 7c), which contains an explanatory variable denoting soil pH, is characterized by a high coefficient of determination (R
2 = 0.6060), although the S
e value should be considered quite high (2.0888). The regression line approximates the real data points quite well for Zn in heavy soil, which can be observed in
Figure 7d, and the model can explain around 84% of the Zn content variability in the soil. Judging by the S
e value, when predicting the value of the dependent variable (Zn), the authors made an error of 0.3164 mg kg
−1 d.m.
Multiple regression analysis resulted in a significant relationship between Mn and Zn availability in very light and heavy soils and some of the explanatory variables (soil pH, S-SO
4 content and dehydrogenase activity). The results are also presented using a Pareto chart of t-values for the regression coefficients (
Figure 8). The Pareto chart shows the size of the effect that each variable involved in the regression model has on the dependent variable (Mn and Zn) in decreasing order, and the line going across the columns indicates how large an effect has to be (i.e., how long a column must be) to be statistically significant. The columns of the plot are in proportion to the calculated values of the t-statistics of each effect. As observed, soil pH was a predictive variable for Zn availability in light and heavy soils as well as for Mn content in heavy soil. The sulfate sulfur content itself seemed to also be an important predictor for Zn content in heavy soil, but less significant than the pH value. In this case, higher values of the t-score for pH in comparison to the S-SO
4 content indicate that a large difference exists between these two variable inputs (approximately 3-fold). The availability of Mn in very light soil significantly depended mostly on dehydrogenase activity.
3.5. Principal Component Analysis (PCA) Biplots to Show the Distribution of Experimental Treatments
Principal component analysis (PCA) was used to display patterns in the data set containing observations described by the quantitative parameters of soil samples (
Figure 10). The PC1 and PC2 components for parameters obtained for very light soil had eigenvalues equal to 3.23 and 0.80, while for heavy soil, these eigenvalues were 3.62 and 0.69, respectively. PC1 and PC2 accounted for 64.51% and 16.07% of the observed variation for very light soil. The amount of variance retained by the principal components (PC1 and PC2) for heavy soil was 72.37% and 13.85%, respectively. For both soil types, the separation of the treatments was shown, but data clustering was more related to incubation than to the experimental treatments themselves.
In the score plot, the data obtained on incubation days 0, 15 and 30 for very light soil were clustered on the positive side of the PC1 axis, with the exception of the SII + D treatment on the 15th and 30th incubation days (negative sides of both PC1 and PC2 axes) (
Figure 10a). Most data obtained on the 30th incubation day were clustered in the lower right quadrant (negative PC2 values), while for incubation day 0, the data were clustered in the upper right quadrant (positive PC2 values), with the exception of SI and SI + B. The negative side of PC1 comprised data measured for treatments on the 60th and 120th incubation days (an exception is C 60). Most of the treatment data collected on the 120th day were grouped in the upper left quadrant of the score plot, especially for the SI and C combination. The highest values for horizontal coordinates were recorded for treatments: SII + M 0 and SI + D 0 (positive) and SII + D 120 (negative). Extreme values on the vertical axis were recorded for treatments: SI + M 0 (positive) and SII 30 (negative). The parameters clustering on the positive side of PC1 were pH and dehydrogenase activity (loadings: 0.86 and 0.76, respectively), and those clustering on the negative side were Zn, Mn and S-SO
4 (loadings: −0.90, −0.79 and −0.70, respectively). PC2 had high positive loadings of Mn (0.58), pH (0.39), Zn (0.35) and dehydrogenase activity (0.26), and the only negative contributor to PC2 was S-SO
4 (−0.35). The first principal component mainly represents variables such as Zn availability in samples of very light soil and soil pH, while the second principal component mainly represents Mn availability.
The data for heavy soil measured on incubation days 0 and 15 were clustered on the positive side of PC1, while data on the 60th and 120th incubation days were located on the negative side of PC1 (
Figure 10b). The data on the 30th day were distributed in all plot quadrants. Almost all points for 0-day treatments were located in the upper right quadrant, with the exception of SII + D 0, which had negative coordinates for PC2. Negative PC1 but positive PC2 coordinates were observed for the parameters measured on the 120th incubation day for all treatments. Parameters measured on the 15th day of treatments with manure (M) and digestate (D) for both sulfur doses were grouped in the bottom right quadrant of the plot. On the other hand, treatments with the double sulfur dose, based on data from the 30th and 60th days, were clustered in the bottom left quadrant. Extreme values on the horizontal axis were recorded for treatments: SII + M 0 and SII + D 0 (positive) and SII 120 (negative). Extreme values on the vertical axis were recorded for treatments: C 120 and SI 120 (positive) and SI + D 30 (negative). The parameters clustering on the positive side of PC1 were soil pH and dehydrogenase activity (loadings: 0.95 and 0.72, respectively), whereas the loadings of Zn, Mn and S-SO
4 were negative (−0.95, −0.75 and −0.74). A positive loading for PC2 was observed only for Mn, but it was very low, below 0.01. Other soil parameters (dehydrogenase activity, S-SO
4, Zn and soil pH) clustered on the negative side of PC2 with loadings of −0.60, −0.56, −0.12 and −0.08, respectively. The first principal component mainly represents variables such as Zn availability in the samples of heavy soil and soil pH, while the second principal component mainly represents dehydrogenase activity and S-SO
4 content.
3.6. Comments on the Results of Advanced Statistical Analyses
The presented findings indicate significant relationships between the soil pH value, S-SO
4, Zn and Mn contents, dehydrogenase activity and experimental duration. The results of the regression analysis revealed that the applied materials had a more significant influence on the zinc content than on the manganese content during the incubation period. The present study used multiple linear regression analysis to identify which soil parameters (pH value, S-SO
4 content and dehydrogenase activity) most significantly affect Mn and Zn contents. This technique is widely used in soil research [
61,
62,
63,
64]. Most of the regression equations after the backward elimination procedure regarded soil pH as a key variable affecting the availability of Zn and Mn contents in both light and heavy soils. This was also confirmed using a Pareto chart of t-values for the most significant regression coefficients in the developed regression models. As presented in the heatmaps, with the prolonging of the incubation period, increasing contents of sulfate sulfur and decreasing values of pH and dehydrogenase activity, the Zn and Mn contents in both tested soils increased. Data clustering in the principal component analysis was more related to incubation duration than the experimental treatments.
The presented findings are in line with our previous research [
65]. From the mentioned experiment, after the preparation of the PCA ordination including soil parameters and conditions for conducting the experiment, it was concluded that the S
0 dose and incubation duration could decrease the soil enzyme activity (dehydrogenase and catalase) and pH value and also increase the content of sulfate sulfur. Furthermore, the liming treatment reduced the relationship strength between the discussed parameters. However, in the case of the elemental sulfur dose and sulfate sulfur content, the lime introduction increased the correlation between these variables. The addition of calcium compounds such as calcium carbonate to low-pH soil increases its oxidative capacity, hence the oxidation of elemental sulfur [
66]. Dall’Orsoletta et al. [
67] reported that the application of elemental sulfur at increasing doses decreases soil pH. As Kulczycki [
68] stated, after studying the effect of the discussed amendment in different doses and types of soils, decreasing soil pH impairs the oxidation rate of the introduced sulfur. Similar results were presented by Zhao et al. [
69], who highlighted a positive correlation between these parameters. A close relationship between the sulfate sulfur content and pH results from the fact that the soil reaction shapes both the abundance and activity of sulfur oxidizers [
49,
70]. These species transform unavailable elemental sulfur form into mobile ions, and their efficiency decreases together with decreasing soil pH values [
70,
71]. A reduction in dehydrogenase activity over the incubation study period was presented by Mierzwa-Hersztek et al., as well [
72]. In addition to decreasing the soil pH value, lowering the value of the discussed parameter could be a result of the scarcity of easily degradable carbon substrates, as well as nutrient resources [
73,
74,
75]. As mentioned in the above section, lower soil pH benefits the mobility of trace elements and reduces the mobility of macroelements. This is a result of the consequences induced by the increased activity of hydrogen ions (H
+). Various relationships among soil parameters have been presented across research reports. Wang et al. [
76], after testing the effect of six different pH values on element concentrations in two soils, reported that, together with the decreasing value of soil pH, the concentrations of plant-available Al, Cd, Mn and Zn increased, while plant-available Ca and Mg contents decreased. According to this, lowering the pH value significantly influenced plant Zn and Cd uptake. In addition, Soaud et al. [
46] highlighted the relationship between an increase in the nutrient concentration and a decrease in the pH value of the tested soils after conducting a 128-day incubation experiment and amending the soil with elemental sulfur. In contrast, Matos Castañon et al. [
77], after conducting a pot experiment, concluded that increasing doses of elemental sulfur did not affect the soil abundance of available P, K and Mg, in spite of lowering the pH of the examined soil. Hanousek et al. [
78], after testing three study sites, found a significant negative correlation between soil pH and the sulfate sulfur content, as well as a significant positive correlation between soil pH and the Al content. Mattiello et al. [
53] found that sulfur oxidation was not correlated with available Zn in the acid soil, contrary to high-pH soil, where it was. Additionally, among the two tested soils, acidic soil was characterized by a higher concentration of this element compared to slightly alkaline soil.